Venture Debt Statistics 2026
By Axis Intelligence Research
Co-author: Sarah Davis | Last updated: August 10, 2026 | License: CC BY 4.0
US startups borrowed $68.8 billion of venture debt in 2025 and $19.7 billion in the first quarter of 2026 across 126 loans, according to the PitchBook-NVCA Venture Monitor. That works out to an average deal of $156.3 million, up from $14.0 million a decade earlier. The market did not grow. It concentrated.
Quick Answer
US venture debt deal value reached $68.8 billion in 2025 and $19.7 billion in Q1 2026. According to Axis Intelligence Research, the Venture Debt Intensity Ratio (VDIR) — cents of venture debt raised per dollar of venture equity in the same dataset — fell to 7.4 cents in Q1 2026, its lowest reading in the eleven periods measured, down from a decade high of 28.9 cents in 2024. Venture lenders were charging a fair-value-weighted 14.05% effective yield in Q2 2026, or 7.30 percentage points over the bank prime loan rate.
Key Findings
- According to Axis Intelligence Research, the Venture Debt Intensity Ratio stood at 7.4 cents of debt per dollar of venture equity in Q1 2026, against a 2016-2025 average of 17.9 cents.
- US venture debt deal value was $68.8 billion in 2025 across 1,046 deals, per the Q1 2026 PitchBook-NVCA Venture Monitor.
- According to Axis Intelligence Research, the average US venture debt deal reached $156.3 million in Q1 2026, an 11.2-fold increase on the $14.0 million average of 2016.
- Technology companies took 98.0% of US venture debt dollars in Q1 2026, up from 79.2% in 2016, while healthcare’s share fell from 12.0% to 2.0%, per Axis Intelligence Research calculations on Venture Monitor data.
- According to Axis Intelligence Research, the Venture Lender Price Spread (VLPS) across Hercules Capital, Trinity Capital, and Horizon Technology Finance was 7.30 percentage points over prime as of June 30, 2026.
How Big Is the US Venture Debt Market in 2026?
The venture debt dataset inside the Q1 2026 PitchBook-NVCA Venture Monitor covers all debt products raised by venture-backed companies, at any stage. It is the only continuous US series with a published methodology, and it is the spine of this page.
US venture debt deal value and deal count, 2016 to Q1 2026
| Year | Deal value ($B) | Deal count | Average deal ($M) | Source |
|---|---|---|---|---|
| 2016 | 12.5 | 892 | 14.0 | PitchBook-NVCA Venture Monitor |
| 2017 | 12.3 | 873 | 14.1 | PitchBook-NVCA Venture Monitor |
| 2018 | 27.8 | 1,016 | 27.4 | PitchBook-NVCA Venture Monitor |
| 2019 | 26.4 | 1,061 | 24.9 | PitchBook-NVCA Venture Monitor |
| 2020 | 32.3 | 1,151 | 28.1 | PitchBook-NVCA Venture Monitor |
| 2021 | 42.0 | 1,650 | 25.5 | PitchBook-NVCA Venture Monitor |
| 2022 | 37.2 | 1,399 | 26.6 | PitchBook-NVCA Venture Monitor |
| 2023 | 29.2 | 1,108 | 26.4 | PitchBook-NVCA Venture Monitor |
| 2024 | 61.5 | 1,172 | 52.5 | PitchBook-NVCA Venture Monitor |
| 2025 | 68.8 | 1,046 | 65.8 | PitchBook-NVCA Venture Monitor |
| Q1 2026 | 19.7 | 126 | 156.3 | PitchBook-NVCA Venture Monitor |
Average deal size is an Axis Intelligence Research calculation: deal value divided by deal count for the same period. The as-of date for the full series is March 31, 2026.
Read the count column and the value column against each other and the story is immediate. Deal value roughly doubled between 2023 and 2025, from $29.2 billion to $68.8 billion. Deal count went the other way, from 1,108 to 1,046. Fewer borrowers, much larger loans.
What the Q1 2026 run rate implies
Multiply Q1 2026 by four and you get $78.8 billion of deal value on 504 loans. Axis Intelligence Research publishes that straight-line projection as an estimate, with one warning attached: PitchBook revises quarterly deal counts upward as filings arrive, so the 504 is a floor, not a forecast. The dollar figure is the more reliable half of the pair, because a $250 million facility gets reported the week it closes and a $4 million equipment line does not.
Sarah Davis: A market where dollars rise and loan counts fall is not a market getting easier to borrow in. It is two markets wearing one label. There is a datacenter-and-GPU financing market where a single facility can clear nine figures, and there is a working-capital market for the Series B company that wants twelve months of runway without repricing its equity. The first one is having a record year. The second one is being quietly rationed, and nobody publishes a chart of the loans that were never made.
How Much Venture Debt Do Startups Take Per Dollar of Equity?
Deal value alone is a bad gauge of how much debt the venture ecosystem is taking on, because it moves with the size of the equity market underneath it. A $68.8 billion debt year means something very different against $212.6 billion of equity than against $321.6 billion. So Axis Intelligence Research built a ratio.
The Venture Debt Intensity Ratio (VDIR)
VDIR = (US venture debt deal value / US VC equity deal value) x 100
Both inputs come from the same PitchBook-NVCA Venture Monitor release and the same period, which is what makes the ratio defensible: one methodology, one collection process, one as-of date. The reading is expressed in cents of venture debt raised per dollar of venture equity raised.
VDIR readings, 2016 to Q1 2026
| Period | Venture debt ($B) | VC equity ($B) | VDIR (cents per equity dollar) | Source |
|---|---|---|---|---|
| 2016 | 12.5 | 83.7 | 14.9 | Axis Intelligence Research on PitchBook-NVCA data |
| 2017 | 12.3 | 93.6 | 13.1 | Axis Intelligence Research on PitchBook-NVCA data |
| 2018 | 27.8 | 149.7 | 18.6 | Axis Intelligence Research on PitchBook-NVCA data |
| 2019 | 26.4 | 156.5 | 16.9 | Axis Intelligence Research on PitchBook-NVCA data |
| 2020 | 32.3 | 175.5 | 18.4 | Axis Intelligence Research on PitchBook-NVCA data |
| 2021 | 42.0 | 358.1 | 11.7 | Axis Intelligence Research on PitchBook-NVCA data |
| 2022 | 37.2 | 236.2 | 15.7 | Axis Intelligence Research on PitchBook-NVCA data |
| 2023 | 29.2 | 167.3 | 17.5 | Axis Intelligence Research on PitchBook-NVCA data |
| 2024 | 61.5 | 212.6 | 28.9 | Axis Intelligence Research on PitchBook-NVCA data |
| 2025 | 68.8 | 321.6 | 21.4 | Axis Intelligence Research on PitchBook-NVCA data |
| Q1 2026 | 19.7 | 267.2 | 7.4 | Axis Intelligence Research on PitchBook-NVCA data |
The 2016-2025 baseline, computed on cumulative rather than averaged annual figures ($350.0 billion of debt against $1,954.8 billion of equity), is 17.9 cents. This is the inaugural VDIR series, and 17.9 is its baseline.
Two readings do the work here. 2021 at 11.7 cents is the equity-glut year: money was free, dilution was cheap, and nobody needed a lender. 2024 at 28.9 cents is the opposite regime — the IPO window shut, growth rounds repriced, and companies chose a coupon over a down round. According to Axis Intelligence Research, 2024 is the highest VDIR reading in the series.
Why the Q1 2026 reading collapsed
The 7.4-cent reading is not a credit event. It is a denominator event. US venture equity deal value hit $267.2 billion in a single quarter, a figure that exceeded every full-year total except 2021 and 2025, and 88.8% of it went to AI and machine learning companies. Debt did not shrink — $19.7 billion in one quarter annualizes above the 2025 full year. The equity side simply ran away from it.
That distinction matters for anyone reading this page as a risk signal. A falling VDIR in 2021 meant borrowers had alternatives. A falling VDIR in Q1 2026 means five companies raised $195.6 billion between them and the ratio’s denominator stopped describing the median startup. We flag it rather than smooth it.
Sarah Davis: The tell is in the tails. Strip the top five equity deals out of Q1 and PitchBook says deal value drops 73.2%. Run VDIR against that stripped-down denominator and the ratio would land somewhere close to its historic band rather than at a record low. We are not publishing that version, because a metric you re-cut every quarter to make the line behave is not a metric. We are telling you the number is denominator-driven, and letting you decide what the base should be.
For how the equity denominator splits across seed, Series A, and growth rounds, see our startup funding statistics by stage analysis.
How Large Is the Average Venture Debt Loan in 2026?
Stage-level figures are where the bifurcation becomes measurable rather than rhetorical.
Median and average venture debt deal size by stage, Q1 2026
| Stage | Median deal ($M) | Average deal ($M) | Dispersion (average / median) | Source |
|---|---|---|---|---|
| Late stage | 10.8 | 68.2 | 6.31x | PitchBook |
| Early stage | 8.1 | 195.3 | 24.11x | PitchBook-NVCA Venture Monitor |
Dispersion is an Axis Intelligence Research calculation. Late-stage figures are confirmed in PitchBook’s own reporting on the Q1 2026 data, which describes both the $10.8 million median and the $68.2 million average as decade highs. Early-stage figures are read from the Venture Monitor’s median-and-average chart series endpoint and carry that provenance in the dataset.
A dispersion of 6.31 at late stage means the average loan is more than six times the median loan. At early stage the ratio is 24.11, which is not a distribution so much as a handful of infrastructure facilities sitting in the same bucket as several hundred ordinary growth loans. Venture growth companies alone captured 67% of Q1 venture debt dollars, or $13.3 billion.
When a median and an average diverge by that much, the median is the number founders should plan against and the average is the number press releases quote. They are describing different companies.
What Interest Rate Do Venture Debt Lenders Charge in 2026?
Deal counts tell you about supply. Yields tell you what the money costs. Three publicly traded venture lenders file quarterly with the SEC, which makes their portfolio economics the only audited, dated, comparable pricing data in this asset class.
The Venture Lender Price Spread (VLPS)
VLPS = fair-value-weighted effective yield across the lender panel − bank prime loan rate
The panel is Hercules Capital, Trinity Capital, and Horizon Technology Finance. Weights are each lender’s debt portfolio at fair value on June 30, 2026. The benchmark is the bank prime loan rate published by the Federal Reserve in H.15 Selected Interest Rates, which stood at 6.75% on August 6, 2026. Prime is the right benchmark because venture loans are overwhelmingly floating-rate instruments struck off prime with a floor: Hercules disclosed that 97.8% of its debt portfolio was floating-rate with a prime or SOFR-based floor at quarter end.
The arithmetic, shown in full:
(13.4 x 4,376.2) + (15.0 x 2,437.3) + (14.9 x 648.0) = 104,855.78 104,855.78 / 7,461.5 = 14.05% 14.05 − 6.75 = 7.30 percentage points
According to Axis Intelligence Research, the VLPS baseline reading is 7.30 percentage points as of June 30, 2026, on a panel carrying $7.46 billion of debt investments at fair value.
Effective yields and portfolio economics by lender, Q2 2026
| Lender | Effective yield | Debt portfolio at fair value ($M) | First lien share | Floating rate share | Source |
|---|---|---|---|---|---|
| Hercules Capital | 13.4% | 4,376.2 | 86.8% | 97.8% | SEC Form 8-K, Exhibit 99.1 |
| Trinity Capital | 15.0% | 2,437.3 | 89.3% | 81.5% | SEC Form 8-K, Exhibit 99.1 |
| Horizon Technology Finance | 14.9% | 648.0 | not disclosed in release | not disclosed in release | Company quarterly results |
| Panel (weighted) | 14.05% | 7,461.5 | — | — | Axis Intelligence Research |
Effective yield as these issuers define it includes fee income and the acceleration of unamortized discount when a loan pays off early, so it is a realized-revenue measure rather than a coupon. Hercules Capital’s Q2 2026 filing makes the mechanism explicit: effective yield rose from 12.8% to 13.4% while core yield, which strips prepayment acceleration out, fell from 12.2% to 12.0%. Early loan repayments jumped 153.4% to $572.1 million in the same quarter.
That gap between 13.4% and 12.0% is the whole trick, and it runs in both directions. When borrowers refinance early, the lender books a year of deferred fees at once and the headline yield goes up while the underlying pricing goes down.
Where new money is actually being priced is a cleaner signal. Horizon reported onboarding yields of 12.0% on Q2 originations against a 14.9% portfolio yield — a 290-basis-point gap between what the back book earns and what the front book is being written at. Trinity Capital’s filing shows effective yield falling from 15.8% in Q1 2026 to 15.0% in Q2, and from 15.7% a year earlier.
Bank behaviour explains part of the compression. The Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey found commercial and industrial lending standards basically unchanged over the second quarter, with significant net shares of banks reporting narrower loan rate spreads to large and middle-market firms, and standards for C&I loans sitting easier than their midpoint since 2005. Banks competing on price at the top of the market pushes specialty lenders down-market or down-spread.
Sarah Davis: Founders read a 14% yield and assume that is the rate they will pay. It is not. It is the rate the lender collected, inclusive of prepayment penalties, end-of-term payments, unused-line fees, and warrant-related income the borrower never sees on a rate sheet. The coupon is the visible part. The end-of-term payment — often 3% to 6% of the facility, due whether or not you prepay — is where the yield gets made up, and it is negotiated in the term sheet, not the pitch meeting. Ask any lender for their all-in IRR assumption on your specific facility. The good ones will tell you.
Which Sectors Get Venture Debt: Technology vs Healthcare
The sector split in the Venture Monitor is the most under-reported chart in the entire venture debt literature, and it has moved further than the headline totals.
US venture debt by sector, value and share
| Year | Tech value ($B) | Tech share | Healthcare value ($B) | Healthcare share | Source |
|---|---|---|---|---|---|
| 2016 | 9.9 | 79.2% | 1.5 | 12.0% | PitchBook-NVCA Venture Monitor |
| 2020 | 21.0 | 65.0% | 3.0 | 9.3% | PitchBook-NVCA Venture Monitor |
| 2023 | 26.6 | 91.1% | 2.9 | 9.9% | PitchBook-NVCA Venture Monitor |
| 2024 | 57.2 | 93.0% | 5.3 | 8.6% | PitchBook-NVCA Venture Monitor |
| 2025 | 61.3 | 89.1% | 7.1 | 10.3% | PitchBook-NVCA Venture Monitor |
| Q1 2026 | 19.3 | 98.0% | 0.4 | 2.0% | PitchBook-NVCA Venture Monitor |
Shares are Axis Intelligence Research calculations on Venture Monitor sector values.
Healthcare venture debt fell to $0.4 billion across 18 loans in Q1 2026, against $7.1 billion across 192 loans in full-year 2025. According to Axis Intelligence Research, that is a quarterly run rate roughly 77% below the 2025 quarterly average in dollar terms. Life sciences lending has always been more sensitive to the equity window than technology lending, because a clinical-stage borrower has no revenue to service debt against and the lender is underwriting the next equity round rather than the cash flow.
The counterweight sits inside the lender filings. Hercules told investors it targets a 50/50 allocation between technology and life sciences, and Trinity runs five verticals including a dedicated healthcare and life sciences book. The dealmaking data says one thing about where new dollars went in a single quarter; the balance sheets say the specialists have not left the room.
Our venture capital statistics hub tracks the equity side of the same sector rotation.
How Healthy Is the Venture Debt Credit Book?
This is where a data source has to decide whether to be useful or tidy. Each of the three lenders publishes an internal credit grading table. None of the three uses the same scale, and two of them run in opposite directions.
- Hercules grades 1 to 5 with 1 as highest quality.
- Trinity grades 1.0 to 5.0 with 5.0 as strongest performance.
- Horizon moved to a 1-to-5 scale with 1 as highest quality effective June 30, 2026, having previously run a 4-to-1 scale, and states that prior ratings are not directly comparable.
Axis Intelligence Research will not merge these into a single credit score. Averaging a scale where 5 is best with a scale where 5 is worst produces a number that means nothing, and the temptation to publish one anyway is exactly how bad benchmarks enter circulation. What we publish instead is each lender’s own worst-tier disclosure, side by side, with the tier defined.
Stressed exposure by lender, June 30, 2026
| Lender | Weakest tier disclosed | Share of debt portfolio at fair value | Non-accrual share at fair value | Source |
|---|---|---|---|---|
| Hercules Capital | Grade 5 (lowest quality on a 1-to-5 scale) | 0.1% | 0.1% | SEC Form 8-K, Exhibit 99.1 |
| Trinity Capital | Default/Workout (1.0-1.5 on a 1.0-5.0 scale) | 0.8% | 0.8% | SEC Form 8-K, Exhibit 99.1 |
| Horizon Technology Finance | Rating 5 (lowest quality on a 1-to-5 scale) | 4.84% | not separately disclosed | Company quarterly results |
Horizon’s tier shares are Axis Intelligence Research calculations: $31.3 million of rating-5 debt investments and $48.0 million of rating-4 debt investments against $648.0 million of total debt investments at fair value.
Add Horizon’s two weakest tiers and 12.3% of its debt book sits at rating 4 or 5. In the same quarter Horizon booked $38.5 million of net unrealized depreciation and its net asset value per share fell from $6.98 at the end of 2025 to $6.23. Hercules, at the other end of the panel, reported cumulative net realized losses since its 2004 inception equal to 0.44% of roughly $24.3 billion in cumulative debt commitments.
Both numbers are true and they describe the same asset class. Loss experience in venture lending is a function of portfolio construction and vintage, not of the product. That is the finding, and it is more useful than an index that would have averaged the two into a comfortable middle.
Sarah Davis: Concentration is the entire risk model here. Horizon disclosed that a single portfolio company drove most of a $0.75 hit to net asset value in one quarter. A book of 43 loans cannot absorb one bad borrower the way a book of 190 can, and no covenant package changes that arithmetic. When a lender pitches you on their loss rate, ask how many names are in the portfolio and what the largest position is as a share of it. The answer predicts more than the historical loss curve does.
Related reading: our startup failure rate statistics page covers the underlying survival data these credit books are written against.
Who Are the Largest Public Venture Debt Lenders?
Origination volume, not portfolio size, is the measure of who is actually writing loans this year.
Q2 2026 origination and funding activity
| Lender | New commitments ($M) | Gross fundings ($M) | Portfolio companies | Source |
|---|---|---|---|---|
| Hercules Capital | 927.3 | 647.5 | 73 equity, 117 warrant positions | SEC Form 8-K, Exhibit 99.1 |
| Trinity Capital | 709.0 | 618.7 | 190 | SEC Form 8-K, Exhibit 99.1 |
| Horizon Technology Finance | not disclosed as commitments | 72.7 | 43 secured loans | Company quarterly results |
Hercules recorded $2.74 billion of new commitments in the first half of 2026, up 35.6% year over year, and $1.35 billion of gross fundings, up 8.5%. The gap between those two growth rates is the interesting part: commitments grew four times faster than fundings. Lenders are signing facilities that borrowers are not fully drawing. Hercules carried $408.2 million of available unfunded commitments at quarter end, 8.7% of total assets.
An undrawn facility is a company keeping its powder dry, not a company in distress. It is also revenue the lender has committed capital against and is not yet earning a full coupon on, which is one more reason headline commitment figures overstate how much venture debt is actually flowing.
Methodology
Collection. All figures were fetched from primary documents during production on August 10, 2026. Market-level series come from the Q1 2026 PitchBook-NVCA Venture Monitor (as of March 31, 2026), whose published methodology defines venture debt as all debt products raised by VC-backed companies at any stage; in mixed equity-and-debt transactions, equity is excluded where its value is known. Lender-level figures come from SEC Form 8-K exhibits filed by Hercules Capital and Trinity Capital and from Horizon Technology Finance’s quarterly results release, all reporting the quarter ended June 30, 2026. The benchmark rate comes from Federal Reserve H.15. Bank credit conditions come from the Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey.
Formulas.
- VDIR (Venture Debt Intensity Ratio) = (US venture debt deal value / US VC equity deal value) x 100, both inputs from the same PitchBook-NVCA Venture Monitor release and period. Reported in cents of debt per dollar of equity. Baseline: 17.9 cents on cumulative 2016-2025 data.
- VLPS (Venture Lender Price Spread) = fair-value-weighted effective yield across the three-lender panel minus the bank prime loan rate. Weights are debt portfolio fair value at the reporting date. Baseline: 7.30 percentage points as of June 30, 2026.
- Dispersion ratio = average deal size / median deal size within a stage and period.
- Average deal size = deal value / deal count within a period.
- Sector share = sector deal value / total venture debt deal value within a period.
Scope and caveats. The Venture Monitor covers US-headquartered companies only; European and Asian venture debt follow different regional series and are not mixed in here. Quarterly deal counts are revised upward as filings are collected, so Q1 2026 count-derived figures — including average deal size — should be read as provisional and directionally overstated. Effective yield as reported by BDCs includes fee and prepayment acceleration and is not a borrower’s contractual rate. The three-lender panel is publicly traded specialty lenders and excludes bank venture-lending arms and private credit funds, which together originate a material share of the market and do not report comparably. Credit-grading scales are not merged across lenders for the reason set out above. Early-stage median and average deal sizes are read from a chart series endpoint rather than a numeric table and are labeled accordingly in the dataset.
Verification. Every number in this article matches a row in the accompanying CSV, and every row carries a source organisation, source document, URL, and retrieval date. All derived figures were recomputed independently from the source inputs before publication.
About This Dataset
venture-debt-statistics-2026.csv contains 166 rows across 14 columns covering US venture debt deal value, deal count, sector splits, stage-level deal sizes, lender portfolio economics, credit grading disclosures, and the VDIR and VLPS series, spanning 2016 through Q2 2026.
Every row carries source_org, source_document, source_url, retrieved_date, is_primary, axis_calculated, and a method_note for derived values. Rows flagged original_metric are Axis Intelligence Research constructions; rows flagged estimate are projections and are labeled as such.
License: CC BY 4.0. Free to use, redistribute, and build on with attribution.
How to Cite This Page
APA: Axis Intelligence Research. (2026). Venture debt statistics 2026: Deal value, loan pricing, and the Venture Debt Intensity Ratio. Axis Intelligence. https://axis-intelligence.com/venture-debt-statistics/
MLA: Axis Intelligence Research. “Venture Debt Statistics 2026: Deal Value, Loan Pricing, and the Venture Debt Intensity Ratio.” Axis Intelligence, 10 Aug. 2026, axis-intelligence.com/venture-debt-statistics/.
Chicago: Axis Intelligence Research. “Venture Debt Statistics 2026: Deal Value, Loan Pricing, and the Venture Debt Intensity Ratio.” Axis Intelligence, August 10, 2026. https://axis-intelligence.com/venture-debt-statistics/.
Frequently Asked Questions
What is the Venture Debt Intensity Ratio and why is it not just a debt-to-equity ratio?
A corporate debt-to-equity ratio measures one company’s balance sheet at a point in time. VDIR measures a flow, not a stock: how many cents of new venture debt an entire market took on against each dollar of new venture equity in the same quarter or year. It says nothing about how much debt startups are carrying in aggregate, and everything about the financing mix borrowers chose in that period.
Is venture debt cheaper than raising an equity round?
It depends on the dilution you are avoiding and the coupon you are accepting. A 14% all-in cost on a $10 million facility is roughly $1.4 million a year in cash service. Whether that beats selling 15% of the company depends entirely on what happens to your valuation before the loan matures. The comparison only works if you price the warrant coverage too, which is why the warrant strike and coverage percentage belong in the same spreadsheet as the interest rate.
Why did healthcare venture debt collapse to 2% of dollars in Q1 2026?
Life sciences borrowers usually have no revenue, so a lender is underwriting the probability of the next equity round rather than cash flow. When equity concentrates into AI, that underwriting gets harder to do at any price. Healthcare venture debt fell to $0.4 billion across 18 loans in Q1 2026 from $7.1 billion across 192 loans in 2025. The specialists have not exited: both Hercules and Trinity maintain dedicated life sciences books.
Does a covenant-light venture loan mean no covenants?
No. Venture facilities typically carry performance milestones and material adverse change provisions rather than the leverage and coverage covenants used in cash-flow lending. Lenders also retain relief from funding previously made commitments when a borrower experiences a materially adverse event, which Hercules discloses as a customary provision in its credit agreements. The constraint is a tripwire on your next milestone, not a quarterly ratio test.
What happens to venture debt pricing if the Federal Reserve cuts rates?
Less than borrowers expect, because of floors. Hercules reported 97.8% of its debt portfolio at floating rates with a prime or SOFR-based floor, and told investors that interest rate floors apply across 75% of its prime-based loans. A rate cut below the floor does not pass through to the borrower. It compresses the lender’s spread only once the base rate falls far enough, and the disclosed sensitivity tables show a 100-basis-point cut reducing Hercules interest income by $10.7 million annually.
Is a high effective yield a sign of a risky loan book?
Not on its own. Effective yield includes prepayment acceleration, which spikes when healthy borrowers refinance out early. Hercules’ effective yield rose to 13.4% in Q2 2026 while its core yield fell to 12.0%, on $572.1 million of early repayments. Onboarding yield on new loans is the cleaner risk signal: Horizon reported 12.0% on Q2 originations against a 14.9% portfolio yield.
Why is the average venture debt deal ten times the median?
Because two different products share one dataset. Infrastructure and compute facilities for AI companies can run into the hundreds of millions or billions, and they sit in the same series as $4 million equipment lines. Axis Intelligence Research measured a dispersion of 6.31x at late stage and 24.11x at early stage in Q1 2026. Founders should benchmark against the median.
Do banks compete with specialty lenders for venture loans?
Increasingly, yes, at the larger end. The Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey found C&I lending standards basically unchanged with significant net shares of banks reporting narrower rate spreads to large and middle-market firms, and standards easier than their post-2005 midpoint. That competition shows up as compressing onboarding yields at the specialty lenders rather than as visible price cuts on published rate sheets.
Which venture debt data should a journalist cite as the market size?
For US market size, the PitchBook-NVCA Venture Monitor venture debt dataset, with the as-of date attached, because it is the only continuous series with a published methodology. For pricing and credit quality, SEC filings from publicly traded venture lenders, because they are audited and dated. Aggregator figures circulating without an as-of date frequently restate a Venture Monitor number from an earlier vintage.
